 So I actually wondered what you think is the biggest real barrier to releasing data from an institutional review board perspective, especially in studies across sites? Well, I think from an IRB perspective it would be a concern about identification of individuals. And when you're dealing with a study that collected samples from locations all around the world, that may not be such a big concern. But when you're dealing with a study that's located in a particular community such as Framingham, you know, a 14-year-old computer geek could pretty reasonably identify a fair number of study participants from the data that we have in our files. You can go to the town hall in Framingham and for $50 purchase a copy of the town census that gives you people's names, their telephone numbers, and their address. You can get their date of birth and you could try to align date of birth of people who share the same last name with the pedigree files that we have in the Framingham database. So the big concern on the part of an IRB is whether we are sufficiently protecting participants. A number of times we've heard about the VA laptop computer that was stolen with a lot of people's data. In fact, I got a letter from the VA saying that my name was among the names of individuals, physicians whose data was stolen as part of that event. So I think that would be the greatest concern. Do you have an answer to that? No, no, no. So my turn to ask you one very quickly and that is in our attempt to identify collaborators around the world for replication of our genome-wide results, we found some studies that seemed to be very ready and prepared to collaborate. But because other studies are aware that Framingham data will go online on October 1st, some of them seem to have a disincentive to want to share with us because it may mean that now they have to share authorships and things of that sort with our investigative team. There's one study I won't name that happens to be in Iceland and another study that I won't name that happens to be in I believe it's Augsburg, Germany that seems to be quite restrictive about the idea of sharing and collaborating. Yeah, I think there are always going to be groups who will take advantage of you releasing your data or making your data available. And I actually don't think that if someone is of that mindset there is anything you can do about it and perhaps you don't particularly want to collaborate with that kind of person. But for every person who does that, I think, or for every group that does that, I think there are many groups who want to truly collaborate and get your intellectual input on pooling of data and analysis across studies. Great. I see we have questions going. Hi. So one of the things that I see changing from the perspective of the consents for the genetic studies, large scale studies in particular, as data were collected in many of the consent forms, is specifically stated, we will not be returning any information to you as a consequence of any genetic studies done on the samples that you provide us with. But now there's an emerging ethical concern that as, you know, if high penetrance polymorphisms are characterized, that there's an obligation to return information to study participants. I was curious whether that came up in any of the discussions with the Framingham participants. Yeah, actually that's a terrific question and you may also be aware of the NHLBI working group that met to address that question. I think Terry was one of the people who organized that. The paper is actually first authored by Ebony Bookman and it would be a useful thing for everyone to be aware of. And it established three conditions for participant notification that there's an association that's identified that's quite strong for disease that carries substantial morbidity and mortality, that we're dealing with a high penetrance situation, so there's a high probability someone will develop that disease. And most importantly, that the knowledge of one's genotype would prompt a treatment that will reduce morbidity and mortality, prevent the development of disease or prevent the complications associated with that disease. And certainly there were a few diseases that were identified and that paper that will fall into that category, one would be the hemopromatosis, allele disease associated allele, where identification of carrier status might result in early treatment to prevent the development of a significant disease. And we approached our Ethics Advisory Board. In fact, our early consent form said we will not attempt to identify you even if we do find an important result. And we've now redone our entire consent form and what I didn't share with you during my presentation was box number eight and box number nine on our consent were permission to notify you and with your permission, your doctor, about genetic conditions you may carry that fall into those categories of notifications. So we agree completely that there are conditions where participants ethically should be notified, in particular when there is an available treatment to prevent a disease or its complications. Yeah, as someone who's spent many years on a tenure decision committee, I detect a little naivete about the importance, or lack of importance of being first or last author on a major paper. So this is not to say that people don't want to share data, but it's a question of when, when it's shared. And so one of the problems with the GAIN project, for example, was I think seen by many as a very short time between well, having short time just to actually write your paper and get it accepted for publication, which I think was initially six months and it might have been extended a little bit. Knowing that there are people with very big resources who are part of the GAIN project out there ready to perhaps work your data a lot more quickly and efficiently than perhaps you could in your small university lab. So the problem is I think in the timing of data release. It's not the fact that it should never obviously be released, but it's just the urgency that some people have been putting on it being released. So if I could summarize, it was a question accusing one of the two of us of naivete. Andy, I think that question was directed at you. Okay, so thanks. So I guess I'll address the second question and then we'll get back to the naivete. I think that one of the review criteria for the GAIN initiative was that there was a solid analytical plan in place. And while I take your point that there are powerful groups out there to analyze the data, I think that if the group that's been reviewed successfully and deposited the samples for the association can't do the core analysis in nine months, then maybe they should let other groups that can do the work get in there and get the real result out. The idea of this work is to find genes associated with disease and we should really strive to do that. On the GAIN initiative, because I was involved in some of the planning of that. And there was a lot of debate as to what an appropriate timeframe was and what the best data release policy would be. And really where we came down, I turn around to face you, Mike, but I'm afraid, I guess this will come with me. Where we came down was that we recognized GAIN was sort of a new effort, a pioneering effort. And we really wanted to kind of push the envelope a little bit. And we have, because now we have the Genes and Environment Initiative and other programs that are putting data out more rapidly. But we recognize that it wouldn't be a good fit for everyone and it hasn't been. But I think you're right, we need to examine multiple models. And actually the way the GAIN works is the data are made available immediately as soon as they're completed and cleaned. But there's a nine month moratorium on submission of publications. So not the actual publication but just submission. In the CGM's data, there's no moratorium. Once the data are put out there, anyone can use them basically. Just getting back to the idea of recognizing the need for people for tenure, to go up for tenure. I mean, I think it's incumbent on those of us in the field who do serve on tenure committees to make sure that the committees understand that science, particularly genetics and epidemiology, is moving towards these kind of large scale efforts. And that we should perhaps shift away from focus of first author publications and last author publications. I realize, I absolutely realize that it's going to take some time for that transition. But the only way it's going to happen is if we push that transition. Well, it is a long way from it. And in the meantime, obviously there are people who are going through the process. So it's a very difficult term. Yeah, I agree. Hi, good afternoon. Larry Kushi from Kaiser Permanente, Northern California. And I wanted to ask a comment about data sharing. Much of the emphasis has been on sharing the results of the genome-wide scans, the 500,000 plus SNPs or whatever. And there was some discussion of sharing the phenotypic data, which typically at least has been described here, is that that's been collected specifically in the context of conducting, say, an epico-word study or whatever. So the questionnaires, et cetera. But in a setting such as I'm in, there's a third stream of phenotypic data, which is the data that comes from clinical and administrative databases that were collected specifically for the care and sort of the tracking of the participants in the health plan, and were not collected specifically in a research study context. And so those data weren't really collected specifically in a consented form, and yet they're available to us as internal researchers. And so those data, probably more so than the data that are collected specifically under research purposes, are constrained by both proprietary interests as well as HIPAA and privacy concerns and sharing of sort of the medical data. And even if they're anonymized, as has been pointed out for some of the genotypic data, there's the potential for being able to link it up and actually identify individuals. And so in that situation, given some of the other challenges like funding of investigators and sort of professional promotion and all that other kind of stuff, I really wonder how these data sharing agreements say if we were to get NIH funding, which we do have, and we extract data from these data sources that were not collected specifically for research purposes but which can be used in a research context, how does that impact what we are able to share? I think that's a big question that we're struggling with internally in terms of how these policies that are coming forth are going to impact our ability to get funding, et cetera. I know time is tight. I'll make a few comments, and Andy may have a couple as well. I think it comes down to the actual language in the consent that participants, that your human subjects, even though they came in for clinical reasons, the Kaiser Permanente, are what they actually signed. So I would recommend that you go back to your IRB with a copy of the consent form. As we did, you may want to form an ethics advisory board that includes some of the members that you're nodding that you actually have done that. And as a third step, you might want to send your documents to a completely outside an involved group of individuals to weigh in as well on their opinion about your ability to perform certain types of research or share data from certain types of research with outside groups. Yeah, I agree. I think getting external advice is key. I mean, I don't think that funding bodies will mandate sharing every single piece of data you have. They only want you to share what's relevant and what can appropriately be shared. I don't think that there's going to be a force to share inappropriate data. Debbie, if this is very quick. It's just really quick. In the Genomics community on sharing, people who share, the more you share, the more information you get back, the more information that you perceive, and people who don't share don't actually get invited anywhere.